On April 11, 1998, about 1845 eastern daylight time, a Starck Auriga, N311RS, registered to an individual, crashed while attempting to land following loss of engine power, at Collegedale Municipal Airport, Collegedale, Tennessee, while on a Title 14 CFR Part 91 personnel flight. Visual meteorological conditions prevailed at the time and no flight plan was filed. The aircraft was destroyed and the commercial-rated pilot received fatal injuries. The flight originated from Collegedale, Tennessee, the same day, about 1843. Witnesses observed N311RS depart on runway 3 at Collegedale Municipal Airport and climb at a steep angle. The pilot then entered a downwind leg for landing on runway 3. Witnesses heard the engine lose power and the pilot transmitted on the airport unicom frequency that he had an engine out. The aircraft was observed to roll to a steep bank angle and enter a base leg, close to the runway. The final leg was over flown and the aircraft was again observed to roll to a steep bank angle in an attempt to line up with the runway. As the aircraft approached the final leg the aircraft rolled rapidly to the left to an inverted position and descended impacting on a road. As the aircraft maneuvered onto the final leg, witnesses heard the engine accelerated momentarily and then stop again. Postcrash examination of the aircraft, engine, and propeller was conducted by an FAA inspector and a representative of Lycoming Engines. The propeller had separated from the engine during impact when the propeller hub separated. Each propeller blade had damage consistent with rotation at the time of ground impact. The engine rotated and continuity of the crankshaft, camshaft, valve train, and accessory drives was established. Each cylinder produced compression. The engine contained a proper quantity of oil and the oil screens did not contain any contamination. Each magneto sparked normally when rotated by hand. The airframe fuel system was reported to have been filled before the flight, taking 41 gallons of fuel. The two fuel tanks, which was formed by the structure in each wing, had ruptured during ground impact. The FAA inspector noted the wing fuel tanks did not have adequate baffling to prevent fuel sloshing and fuel pick up unporting during maneuvering. The fuel selector operated normally and was unobstructed. The engine fuel system contained uncontaminated 100LL fuel. All fuel lines were found unobstructed. The engine fuel system was removed from the engine and tested by NTSB. The engine-driven fuel pump, fuel injector unit, fuel manifold, and each fuel injector operated normally. The electric fuel boost pump switch was found in the on position after the accident. The electric fuel pump did not operate after the accident. The pump was examined, under the supervision of an FAA inspector, by Weldon Pump, Inc., Oakwood Village, Ohio, the manufacturer. The negative power wire had been pulled from the pump motor and was missing. A replacement wire was installed and the motor operated normally. The pump portion of the unit was tested and found to leak from the pump drain. An incorrect O-ring was found installed in the pump. The correct O-ring was installed and the pump operated normally. Weldon Pump personnel stated in their report that the incorrect O-ring installed in the pump would cause the engine-driven fuel pump of the aircraft's engine to draw air into the engine fuel system. (See attached Weldon Pump, Inc. report) Witnesses stated to an FAA inspector that the pilot had experienced an engine failure several weeks before the accident. The pilot had made a forced landing at William's Field, near Collegedale. It was reported that the engine problem was related to failure of the electric fuel boost pump and that the pilot had made repairs to the pump before the accident flight. Postmortem examination of the pilot was performed by F. K. King, Jr., M.D., Hamilton County Medical Examiner's Office, Chattanooga, Tennessee. The cause of death was attributed to multiple injuries and there were no findings that could be considered causal to the accident. Postmortem toxicology studies on specimens obtained from the pilot were performed by University of Tennessee Laboratory, Memphis, Tennessee, and the FAA Toxicology Laboratory, Oklahoma City, Oklahoma. The tests were negative for carbon monoxide, cyanide, ethanol, and basic, acidic, and neutral drugs. The tests were positive for ibuprofen. (See Supplement K and the toxicology reports attached to this report.)